[prev in list] [next in list] [prev in thread] [next in thread] 

List:       theora
Subject:    Re: [theora] NHW Project - high compression -l4 setting
From:       Raphael Canut <nhwcodec () gmail ! com>
Date:       2018-02-19 15:24:56
Message-ID: CAKE58qF6oqLhwnEe9DM9VjGJfQmsBf799v=SY3s1Txicz376tA () mail ! gmail ! com
[Download RAW message or body]

[Attachment #2 (multipart/alternative)]


Hello,

-l4 compression setting is still experimental, I like its
neatness/sharpness approach because the other codecs at equivalent
compression tend to blur image.Do you approve this approach?

On the other side, -l4 setting little lacks of precision, should raise the
quantization from 0.87 to something like 0.89-0.9, and then I would have to
compensate with for example more quantization on the DC parts or start to
increase "adaptively" the wavelet coeff threshold of the dead-zone.

-l3 setting seems to be good, even if I can still raise quantization to
0.89-0.9 and save bits on other parts...

As I said in my last post, when -l3 and especially -l4 settings will be
validated, I will create a new release 0.1.4 for the NHW Project, this is
for long overdue.-Even if there will not be all the improvements I wanted
to add, for example we can still save 1-2Ko in average per .nhw file...-

Many thanks!
Cheers,
Raphael

2018-02-18 15:35 GMT+01:00 Raphael Canut <nhwcodec@gmail.com>:

> Hello,
>
> For those interested, I have added a -l4 high compression setting to the
> NHW codec ( http://nhwcodec.blogspot.com/ ).This setting is very
> experimental, I have boosted the sharpness for this one, but it can lack of
> precision on low contrast areas.
>
> This setting still maintains a good neatness and when it will be
> validated, I will then make a new 0.1.4 version for the NHW Project.
>
> High compression is really complex.Any help is welcome.
>
> Cheers,
> Raphael
>

[Attachment #5 (text/html)]

<div dir="ltr"><div><div><div><div><div><div><div>Hello,<br><br></div>-l4 compression \
setting is still experimental, I like its neatness/sharpness approach because the \
other codecs at equivalent compression tend to blur image.Do you approve this \
approach?<br><br></div>On the other side, -l4 setting little lacks of precision, \
should raise the quantization from 0.87 to something like 0.89-0.9, and then I would \
have to compensate with for example more quantization on the DC parts or start to \
increase &quot;adaptively&quot; the wavelet coeff threshold of the \
dead-zone.<br><br></div>-l3 setting seems to be good, even if I can still raise \
quantization to 0.89-0.9 and save bits on other parts...<br><br></div>As I said in my \
last post, when -l3 and especially -l4 settings will be validated, I will create a \
new release 0.1.4 for the NHW Project, this is for long overdue.-Even if there will \
not be all the improvements I wanted to add, for example we can still save 1-2Ko in \
average per .nhw file...-<br><br></div>Many \
thanks!<br></div>Cheers,<br></div>Raphael<br></div><div class="gmail_extra"><br><div \
class="gmail_quote">2018-02-18 15:35 GMT+01:00 Raphael Canut <span dir="ltr">&lt;<a \
href="mailto:nhwcodec@gmail.com" \
target="_blank">nhwcodec@gmail.com</a>&gt;</span>:<br><blockquote class="gmail_quote" \
style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div \
dir="ltr"><div><div><div><div><div>Hello,<br><br></div>For those interested, I have \
added a -l4 high compression setting to the NHW codec ( <a \
href="http://nhwcodec.blogspot.com/" \
target="_blank">http://nhwcodec.blogspot.com/</a> ).This setting is very \
experimental, I have boosted the sharpness for this one, but it can lack of precision \
on low contrast areas.<br><br></div>This setting still maintains a good neatness and \
when it will be validated, I will then make a new 0.1.4 version for the NHW \
Project.<br><br></div>High compression is really complex.Any help is \
welcome.<br><br></div>Cheers,<br></div>Raphael<br></div> \
</blockquote></div><br></div>


[Attachment #6 (text/plain)]

_______________________________________________
theora mailing list
theora@xiph.org
http://lists.xiph.org/mailman/listinfo/theora


[prev in list] [next in list] [prev in thread] [next in thread]